This project will develop model-based approaches to therapy switching for HIV which minimize the risk of subsequent treatment failures by simultaneously considering the contributions of the viral and proviral genetic makeup and the total viral load to this risk. Using components of previously failed antiviral regimens, the patient's viral load will be "pre-conditioned" for the new antiviral regimen, to maximize its possibility of success. This approach differs from the previous approaches in several important ways. It focuses on therapy switches in patients who have not yet developed multi-drug resistant virus, for whom potentially successful antiviral regimens still exist. It incorporates into its calculation of risk not only the known incidences of cross-resistance between antiviral drugs, but also the knowledge of genetic distance between dominant resistant strains, based on the patient's history of antiviral use. These approaches to treatment will be based on mathematical models of HIV quasispecies dynamics and the competition between various viral strains during different treatment application schedules, as well as mathematical models of resistance emergence. These models will be used to develop laboratory and clinical experiments as future work to implement these techniques. The model- based algorithmic therapy-switching schedules should significantly reduce the risk of failure of subsequent antiviral therapy by reducing the risk of resistant virus pre-existing its introduction. PUBLIC HEALTH RELEVANCE: If successful, this research will provide a method to reduce the incidence of HIV treatment failure due to resistant virus. This has the potential to extend the life expectancy of thousands of people infected with the HIV virus, and to improve their quality of life. Application of this method may also reduce the overall incidence of multi-drug resistant virus in the population.